Preventing Online Fraud Should Start by Examining Customer Intent

There are promises made in commerce — unspoken promises that are too often broken, inadvertently, in the digital age.

Businesses have pledged to deliver high-quality customer experiences. Friction is the ever-present enemy, especially at checkout.

But as more and more transactions go online, opening themselves to online fraud as a result, enterprises must step up their identity verification and authentication. Making those processes a bit more “human,” seamless and more easily navigable are critical to improving conversions.

Neuro-ID CEO Jack Alton and Mission Lane Head of Enterprise Fraud, Collections and Recoveries Gauri Gopalakrishnan told PYMNTS’ Karen Webster in the latest On The Agenda conversation that behavioral analytics that focus on pre-submit data can stop fraud in its tracks.

Companies must put the consumer at the center of the digital experience — but getting there is easier said than done. The bad actors are growing ever more sophisticated in their efforts to infiltrate accounts, pose as legitimate customers, and make off with money or goods.

In the past two years, said Gopalakrishnan, companies — traditional banks and digital startups alike — have had to navigate a rising tide of online attacks while trying to verify new accounts.

As Alton noted, “the intention to deliver a great customer experience oftentimes has to be balanced out by making sure that the identity check you’ve done, as a business, is going to protect the bottom line.”

That’s a high hurdle when it’s getting easier to hide the facts about who you are — or steal them from someone else. Perhaps paradoxically, Alton observed, recent movements by companies like Apple to protect consumers (through new rules tied to app tracking) have made it easier for bad actors to mask email addresses.

The Crossroads

Technology is where the two crossroads meet, where the goal of screening identities with minimal friction — and without getting customers too involved in the process, so that they abandon the transaction — becomes reality. Companies like Mission Lane, he added, can use behavioral analytics tools to determine that someone is really who they say they are.

Gopalakrishnan said that customers — and yes, criminals — can throw off signals that paint a portrait of their intent every time they make a transaction online.

As he explained to Webster, “we try to minimize the instances or the use cases in which we bring that human element, while realizing the fact that there will always be some cases where we will need something ‘extra’” in the form of stepped-up authentication. That could involve something as simple as taking a picture of your driver’s license and sending it to the business you’re looking to transact with.

But before getting to that stage, companies should strive to collect the most data that they can to make sure, at the point of assessment and onboarding, that they get an accurate picture of who’s on the other side.

The Benefits of Pre-Submit Data

Behavioral analytics tools help finance and risk management teams avoid becoming “the show-stoppers that had been getting in the way of conversion, preventing good orders from getting through,” Alton said.

They give companies a crowd-level view of bot attacks and behaviors that look risky — and they can alert the crowds of legitimate users to be vigilant, while letting more “good” commerce in through the front door.

“When you unlock that pre-submit data — how that person actually interacted with you — it can point you toward 10%, 20%, 30%, 40% of your customers that are genuine, that you don’t have to subject to friction,” Alton said.

There’s also good news to be found in the fact that, as Gopalakrishnan observed, it’s easier than ever to measure the damage of false positives — and address the customer experience problems they present straightaway.

“People will tell you upfront about a bad experience,” he said. “They will call in to complain or will complain to regulators. Fraud losses take more time to figure out.”

In part, that slow burn comes because criminals often bide their time, waiting to exploit the right vulnerability. It’s partly because it can take time to uncover the fraud and alert to consumers. Algorithms and machine learning models, said Alton, can help determine what behaviors ultimately wind up leading to good customer journeys, or what behaviors of individuals (and groups) result in fraud.

For many retailers, where online sales conversion rates can be in the low-single-digit percentage points, visibility into customer behavior can lift those rates dramatically. And increasing conversion at scale can approve results, no matter the vertical.

As Alton noted, examining intent is a way of taking “a more holistic view of what my customer’s doing, not just at an individual level,” while taking cues from the broader ecosystem too.

“There is wisdom in the crowd,” he told Webster.